CN109243546A - The method for building up and device of anticoagulation drug effect Optimized model - Google Patents
The method for building up and device of anticoagulation drug effect Optimized model Download PDFInfo
- Publication number
- CN109243546A CN109243546A CN201811045577.9A CN201811045577A CN109243546A CN 109243546 A CN109243546 A CN 109243546A CN 201811045577 A CN201811045577 A CN 201811045577A CN 109243546 A CN109243546 A CN 109243546A
- Authority
- CN
- China
- Prior art keywords
- drug effect
- anticoagulant
- drug
- sample database
- optimization
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000000857 drug effect Effects 0.000 title claims abstract description 116
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000010100 anticoagulation Effects 0.000 title abstract description 8
- 238000005457 optimization Methods 0.000 claims abstract description 154
- 239000003814 drug Substances 0.000 claims abstract description 55
- 229940079593 drug Drugs 0.000 claims abstract description 50
- 238000007689 inspection Methods 0.000 claims abstract description 38
- 230000006870 function Effects 0.000 claims abstract description 25
- 238000012545 processing Methods 0.000 claims abstract description 15
- 239000003146 anticoagulant agent Substances 0.000 claims description 102
- 229940127219 anticoagulant drug Drugs 0.000 claims description 97
- 238000012360 testing method Methods 0.000 claims description 47
- 238000003066 decision tree Methods 0.000 claims description 23
- 230000023555 blood coagulation Effects 0.000 claims description 17
- 238000012216 screening Methods 0.000 claims description 15
- 238000012217 deletion Methods 0.000 claims description 14
- 230000037430 deletion Effects 0.000 claims description 14
- 230000000694 effects Effects 0.000 claims description 12
- 230000003285 pharmacodynamic effect Effects 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 10
- 238000004140 cleaning Methods 0.000 claims description 9
- 230000000361 pesticidal effect Effects 0.000 claims description 9
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000007619 statistical method Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 238000005259 measurement Methods 0.000 claims description 4
- 230000007812 deficiency Effects 0.000 claims description 3
- 238000013433 optimization analysis Methods 0.000 claims description 2
- 238000010801 machine learning Methods 0.000 abstract description 4
- 230000015271 coagulation Effects 0.000 abstract description 3
- 238000005345 coagulation Methods 0.000 abstract description 3
- 238000000528 statistical test Methods 0.000 abstract description 3
- 238000012913 prioritisation Methods 0.000 abstract 1
- 238000011282 treatment Methods 0.000 description 10
- 230000008569 process Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 208000007536 Thrombosis Diseases 0.000 description 3
- 206010053567 Coagulopathies Diseases 0.000 description 2
- 238000000959 Cochran–Mantel–Haenszel (CMH) test Methods 0.000 description 2
- 208000005189 Embolism Diseases 0.000 description 2
- 208000032843 Hemorrhage Diseases 0.000 description 2
- 238000000585 Mann–Whitney U test Methods 0.000 description 2
- 238000003759 clinical diagnosis Methods 0.000 description 2
- 230000035602 clotting Effects 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 108010039209 Blood Coagulation Factors Proteins 0.000 description 1
- 102000015081 Blood Coagulation Factors Human genes 0.000 description 1
- 102000001554 Hemoglobins Human genes 0.000 description 1
- 108010054147 Hemoglobins Proteins 0.000 description 1
- 108010094028 Prothrombin Proteins 0.000 description 1
- 102100027378 Prothrombin Human genes 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000000740 bleeding effect Effects 0.000 description 1
- 239000003114 blood coagulation factor Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000036267 drug metabolism Effects 0.000 description 1
- 238000011221 initial treatment Methods 0.000 description 1
- 238000007477 logistic regression Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000001422 normality test Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 229940039716 prothrombin Drugs 0.000 description 1
- 238000013138 pruning Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000002560 therapeutic procedure Methods 0.000 description 1
- 230000017105 transposition Effects 0.000 description 1
Landscapes
- Investigating Or Analysing Biological Materials (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The method for building up and device of offer anticoagulation drug effect Optimized model of the embodiment of the present invention, wherein method includes: to establish drug effect optimization sample database, and each sample in the drug effect optimization sample database includes Medication order information and the inspection item data corresponding with the Medication order information when coagulation function is normal；Missing values processing and statistical test are carried out to drug effect optimization sample database, obtains and optimizes sample database by pretreated drug effect；Optimize sample database by pretreated drug effect according to described, is exercised supervision study using XGBoost algorithm, building anticoagulation drug effect Optimized model.The embodiment of the present invention constructs anticoagulation drug effect Optimized model using machine learning XGBoost algorithm, the individual instances of different patients can be directed to, rapidly obtain the stronger anticoagulation drug effect prioritization scheme of practicability by the way that the inspection of patient is checked that data input in the anticoagulation drug effect Optimized model.
Description
Technical Field
The invention relates to the technical field of medical artificial intelligence, in particular to a method and a device for establishing an anticoagulant drug effect optimization model.
Background
Anticoagulant drugs are drugs that prevent the coagulation process by affecting some coagulation factors in the coagulation process, and can be used for preventing and treating diseases of intravascular embolism or thrombosis, and preventing stroke or other thrombotic diseases. However, most of anticoagulants have large differences in individual drug metabolism capability, narrow treatment safety range, and many factors influencing the drug effect, so that the individual dose variability of the anticoagulants is large, serious potential bleeding risks and embolism caused by insufficient anticoagulation also exist, and the management of treatment with the anticoagulants is challenging.
Currently, the efficacy optimization scheme of clinical anticoagulants generally adopts a certain standard dosage to be given firstly, and then a clinician repeatedly increases or decreases the dosage according to the INR (International Normalized Ratio) value of each patient until the INR reaches a target. In such anticoagulation therapy, the period for adjusting the dose is long, and the patient is highly likely to suffer from thrombosis or hemorrhage.
In order to overcome the defects of the prior clinical anticoagulant drug effect optimization scheme, "Laboratory Medicine, December 2013, Vol28.No12: 1157-: a stable dose prediction model, a starting dose prediction model, an accurate model of a stable dose, but still not to the extent that individualized dosing is achieved from individual patient to individual patient.
Disclosure of Invention
In order to overcome the defect that the existing anticoagulant medicine medical advice information can not realize individualized medicine application according to different individual patients, the embodiment of the invention provides an establishing method and device of an anticoagulant medicine efficacy optimization model.
In a first aspect, an embodiment of the present invention provides a method for establishing an anticoagulant drug efficacy optimization model, including:
establishing a drug effect optimization sample database, wherein each sample in the drug effect optimization sample database comprises medication advice information when the blood coagulation function is normal and test item data corresponding to the medication advice information;
carrying out deletion value processing and statistical inspection on the pesticide effect optimization sample database to obtain a pretreated pesticide effect optimization sample database;
and carrying out supervised learning by using an XGboost algorithm according to the preprocessed drug effect optimization sample database to construct an anticoagulant drug effect optimization model.
In a second aspect, an embodiment of the present invention provides an apparatus for establishing an anticoagulant drug efficacy optimization model, including:
the system comprises a sample acquisition module, a drug effect optimization sample database and a drug effect optimization analysis module, wherein each sample in the drug effect optimization sample database comprises medication advice information when the blood coagulation function is normal and test item data corresponding to the medication advice information;
the preprocessing module is used for carrying out deletion value processing and statistical inspection on the drug effect optimization sample database to obtain a preprocessed drug effect optimization sample database;
and the model establishing module is used for carrying out supervised learning by utilizing an XGboost algorithm according to the preprocessed drug effect optimization sample database to establish an anticoagulant drug effect optimization model.
In a third aspect, an embodiment of the present invention provides an electronic device, including:
at least one processor; and
at least one memory communicatively coupled to the processor, wherein:
the memory stores program instructions executable by the processor, and the processor calls the program instructions to perform the method provided by any of the various possible implementations of the first aspect.
In a fourth aspect, embodiments of the present invention provide a non-transitory computer-readable storage medium storing computer instructions that enable a computer to perform a method provided in any one of the various possible implementations of the first aspect.
According to the method and the device for establishing the anticoagulant drug effect optimization model, the machine learning XGboost algorithm is adopted to establish the anticoagulant drug effect optimization model, and the anticoagulant drug effect optimization scheme with higher practicability can be quickly obtained by inputting inspection and inspection data of patients into the anticoagulant drug effect optimization model according to individual conditions of different patients.
Drawings
Fig. 1 is a schematic flow chart of a method for establishing an anticoagulant drug effect optimization model according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of an apparatus for establishing an anticoagulant drug effect optimization model according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, a schematic flow chart of a method for establishing an anticoagulant drug efficacy optimization model provided in an embodiment of the present invention includes:
s1, establishing a drug effect optimization sample database, wherein each sample in the drug effect optimization sample database comprises medicine taking advice information when the blood coagulation function is normal and test item data corresponding to the medicine taking advice information.
Specifically, the embodiment of the invention aims to establish an anticoagulant drug efficacy optimization model by adopting a supervised learning method, wherein the anticoagulant drug efficacy optimization model is used for predicting a corresponding drug efficacy optimization scheme according to test item data of a patient. Therefore, sample data needs to be acquired before the supervised learning method is used. Each sample includes two pieces of information, that is, medication order information including information such as a medication dose, a medication interval, and a medication route when a blood coagulation function of a patient returns to normal after the patient uses an anticoagulant, and that is, test item data of the patient corresponding to the medication order information includes, for example: mean hemoglobin amount, platelet count, mean platelet volume, and prothrombin time. In order to achieve this, medication order information in the case of normal blood coagulation function needs to be encoded.
In order to enable the anticoagulant drug efficacy optimization model constructed in the embodiment of the present invention to have higher prediction accuracy, a large amount of patient data, which usually comes from different hospitals, needs to be collected during implementation, so in the embodiment of the present invention, the collected data is stored in a database form, and the data is conveniently processed by adopting the database form.
Specifically, INR values between 2 and 3 indicate normal clotting function.
It should be noted that, according to the needs of different treatment periods, the medication order information is generally divided into initial medication order information and adjustment medication order information, and the construction of the drug effect optimization sample database also needs to be divided accordingly.
The data contained in the drug effect optimization sample database is not directly available for supervised learning, and needs to be preprocessed.
And S2, carrying out missing value processing and statistical inspection on the drug effect optimization sample database to obtain a preprocessed drug effect optimization sample database.
Specifically, the missing value processing refers to deleting the check items with a high data missing rate aiming at the data corresponding to each check item in the pharmacodynamic optimization sample database, so that the validity of the sample can be ensured.
The statistical test is to use a statistical method to test the data corresponding to each test item in the drug effect optimization sample data for data distribution and significance, retain the test item having significant relation with the drug effect of the anticoagulant drug, realize the pre-screening of the test item, and realize the data dimension reduction.
And S3, performing supervised learning by using an XGboost algorithm according to the preprocessed drug effect optimization sample database, and constructing an anticoagulant drug effect optimization model.
The XGboost algorithm is an improved algorithm based on the GBDT (gradient spanning tree) principle, is the fastest and best boost tree algorithm at present, can realize parallel operation and incremental learning, and can process large-scale data.
And extracting sample data from the drug effect optimization sample database to form a data set, and dividing the data set into a training set and a test set.
According to the preprocessed pharmacodynamic optimization sample database, the XGboost algorithm is used for quickly constructing the anticoagulant pharmacodynamic optimization model with high accuracy, and the method specifically comprises the following steps:
inputting a training set, taking each test item as an independent variable, and taking a pesticide effect optimization scheme as a dependent variable;
secondly, defining an objective function, wherein the objective function comprises two parts of loss and regularization terms;
where the loss is the error (gradient) of the last tree and the regularization term is the complexity of the tree. It is desirable to optimize the objective function so that the prediction error of the objective function is as small as possible and the complexity of the numbers is as low as possible.
Thirdly, utilizing a greedy method to search segmentation points and constructing a decision tree;
specifically, all different tree structures may be enumerated, and a scheme with the largest Gain value exceeding a threshold may be selected, and pruning may be terminated if max (Gain) is less than the threshold.
After the decision tree structure is determined, calculating the scores of the leaf nodes;
step five, updating the decision tree sequence, and storing all constructed decision trees and scores thereof;
calculating the prediction result of each sample, namely the sum of scores of each tree, and obtaining the probability that the sample belongs to each category;
step seven, calculating the importance score of each variable, and selecting important variables which have obvious influence on the model;
specifically, Gini (kini) coefficients of each variable may be calculated, and the average value of the Gini coefficients is the importance score of the variable.
And step eight, constructing an anticoagulant drug effect optimization model by using the important variables.
And after the anticoagulant drug effect optimization model is constructed, inputting a test set into the anticoagulant drug effect optimization model for testing, and calculating the prediction accuracy of the test set.
And inputting the test item data to be predicted into the constructed anticoagulant drug effect optimization model, so as to obtain a drug effect optimization scheme corresponding to the test item data to be predicted.
According to the method for establishing the anticoagulant drug efficacy optimization model provided by the embodiment of the invention, the anticoagulant drug efficacy optimization model is established by adopting a machine learning XGboost algorithm, and the inspection and inspection data of the patient are input into the anticoagulant drug efficacy optimization model, so that an anticoagulant drug efficacy optimization scheme with higher practicability can be quickly obtained, and individualized drug efficacy optimization can be realized for different patients.
Further, based on the above embodiment, the step of establishing a drug efficacy optimization sample database specifically includes:
acquiring clinical data of a patient using an anticoagulant and performing data cleaning;
sorting the clinical data of the patients after data cleaning according to time, extracting medication order information with an INR value of 1.5-2.5 after the patients use the anticoagulant drugs for the first time for three days and corresponding test item results from the clinical data of the patients to form an initial scheme database;
and extracting the medication order information when the INR value is not between 2 and 3 and the dosage is adjusted until the blood coagulation function of the patient is recovered to be normal after the first anticoagulant administration for three days and all the test item results after the dosage is adjusted to form an adjustment scheme database.
In particular, the patient clinical data comprises at least: patient physiological characteristic information, clinical diagnosis data, inspection data and medication advice information, wherein the patient physiological characteristic information comprises: height, weight, age, etc.; clinical diagnosis data is related to diseases of patients, the inspection data refers to inspection result data of various examinations capable of influencing blood coagulation function, and medication advice information, namely, a scheme for optimizing the drug effect given by a doctor after analyzing the clinical inspection data of the patients, comprises single administration dosage, administration interval, administration route and the like.
Data cleansing is a process of reviewing and verifying acquired clinical data of patients using anticoagulants, aiming at deleting repeated information, correcting obviously wrong information and ensuring data consistency, and specifically comprises the following steps: data normalization processing, outlier processing, data transposition, data grouping, data deduplication, data sorting, data merging, one-hot encoding, and the like.
According to the requirements of different treatment periods, the medication order information is generally divided into initial medication order information and adjustment medication order information, so that the patient clinical data after data cleaning is sorted according to time, and then the collected patient clinical data is divided according to different treatment periods.
Since INR values between 2 and 3 indicate normal clotting function, the following criteria were used for data partitioning:
and extracting the medication order information of which the INR value is between 1.5 and 2.5 after the anticoagulant is used for the first time for three days and the corresponding test item result to form an initial scheme database.
And (3) providing the medication order information when the INR value is not between 2 and 3 after the first anticoagulant administration for three days and the dosage is adjusted until the blood coagulation function of the patient is recovered to be normal and all the test item results after the dosage is adjusted to form an adjustment scheme database.
The initial scheme database and the adjustment scheme database jointly form a drug effect optimization sample database.
The embodiment of the invention divides the sample data according to the treatment time, so that the method for optimizing the drug effect of the anticoagulant provided by the embodiment of the invention can predict the drug effect optimization scheme which accords with the actual treatment stage.
Further, based on the above embodiment, the step of performing deficiency value processing and statistical test on the pharmacodynamic optimization sample database to obtain a preprocessed pharmacodynamic optimization sample database specifically includes:
deleting the inspection item with the data deletion rate larger than a preset threshold value in the drug effect optimization sample database;
and (3) performing data distribution inspection on the remaining inspection item data by adopting a statistical method, screening out significant variables influencing the drug effect of the anticoagulant drug, and obtaining a preprocessed drug effect optimization sample database.
Specifically, the preset threshold is obtained by the following method: setting the range of the threshold value of the deletion rate to be 50% -95%, taking 5% as an adjusting unit, deleting the inspection items with the deletion rates of more than 50%, 55%, 60%, … … and 95% successively, verifying the accuracy of the test set in sequence, and searching the optimal threshold value of the deletion rate with the highest accuracy of the test set as the preset threshold value.
After deleting the inspection item with the data deletion rate larger than the preset threshold in the drug effect optimization sample database, firstly judging the type of each inspection item aiming at the remaining inspection item data, wherein the type of the inspection item can be divided into continuous variables, namely the result corresponding to the inspection item is a continuous value, and can also be divided into multi-classification variables and two-classification variables.
The data distribution test includes a data bias test and a Mann-Whitney-Wilcoxon test, wherein,
the data bias test is used to test whether the continuous variables conform to normal distribution, and the common statistical method is as follows: normality test, kurtosis, skewness, P-P diagram, Q-Q diagram and the like. If the data does not fit a normal distribution, the data is transformed by 1+ log (x).
Mann-Whitney-Wilcoxon test: is a nonparametric test of two independent samples to check whether data from different hospitals conform to the same distribution. The original assumption is that: data from different hospitals fit the same distribution. When the original hypothesis is accepted, the data quality is considered to be better and can be used for expanding data processing.
The specific process for screening out the significant variables affecting the drug effect of the anticoagulant comprises the following steps:
for the binary variables, performing Cochran-Mantel-Haenszel test on the binary variables and the target variables to judge whether the relation between the binary variables and the target variables is obvious, wherein the original hypothesis of the Cochran-Mantel-Haenszel test is as follows: there is no significant relationship between the binary variables and the target variable. If the original assumption is rejected, the relation between the two classification variables and the target variable is considered to be obvious, the variables are reserved, and otherwise, the variables are removed.
For the multi-classification variable, judging whether the relation between the multi-classification variable and the target variable is obvious or not by performing Cochran-Armitage trend test on the multi-classification variable and the target variable, wherein the Cochran-Armitage trend test is originally assumed to be as follows: there is no significant relationship between the multi-classification variable and the target variable. If the original assumption is rejected, the relation between the multi-classification variable and the target variable is considered to be obvious, the variable is reserved, and if not, the multi-classification variable is eliminated.
And (3) performing logistic regression on the continuous variable and the target variable by adopting a stepwise regression (LR) mode, gradually screening the continuous variable which has obvious influence on the target variable, and rejecting the continuous variable if the continuous variable is not obvious.
The target variable in the above process is the coded medication order information when the blood coagulation function is normal, the total number of the medication order information of the anticoagulant is certain, each medication order information needs to be coded, and the target variable can be regarded as a multi-classification variable.
According to the method for establishing the anticoagulant drug effect optimization model, variables influencing the anticoagulant drug effect are pre-screened before the model is established, so that the establishment speed of the optimization model is improved, and the prediction accuracy of the model is improved.
Further, based on the above embodiment, the step of performing supervised learning by using the XGBoost algorithm according to the preprocessed pharmacodynamic optimization sample database to construct an anticoagulant pharmacodynamic optimization model specifically includes:
initializing XGboost algorithm parameters, wherein the algorithm parameters comprise: maximum decision tree number, learning rate, maximum rule layer depth, minimum gain value required by decision tree growth and decision tree complexity measurement parameters;
based on the initial scheme database, screening out important variables influencing the initial drug effect of the anticoagulant drug by using an XGboost algorithm, and constructing an initial drug effect optimization model according to the important variables influencing the initial drug effect of the anticoagulant drug;
based on the adjustment scheme database, screening out important variables influencing the adjustment effect of the anticoagulant drug by using an XGboost algorithm, and constructing an adjustment effect optimization model according to the important variables influencing the adjustment effect of the anticoagulant drug;
and combining the initial efficacy optimization model and the adjusted efficacy optimization model to generate an anticoagulant efficacy optimization model.
In one embodiment, the XGBoost algorithm parameters may be configured to: the maximum number of decision trees is 2000; learning rate is 0.01; maximum regular layer depth is 4; the minimum Gain value required by the growth of the decision tree is 0; the decision tree complexity measure parameter is 1.
Inputting the preprocessed initial scheme database into a configured XGboost algorithm model, defining an objective function comprising loss and regularization terms, utilizing a greedy method to search for segmentation points, constructing decision trees, calculating scores of leaf nodes, updating decision tree sequences, storing all constructed decision trees and scores thereof, calculating a prediction result of each sample, namely the sum of the scores of each tree, obtaining the probability that the sample belongs to each category, calculating the importance score of each variable, selecting important variables which have obvious influence on the initial efficacy of the anticoagulant drug, and obtaining the decision tree sequences according to the important variables which influence the initial efficacy of the anticoagulant drug, thereby obtaining an initial efficacy optimization model.
And inputting the preprocessed adjustment scheme database into a configured XGboost algorithm model, and finally obtaining an adjustment efficacy optimization model by adopting the same construction process.
And combining the initial efficacy optimization model and the adjusted efficacy optimization model to obtain a trained anticoagulant efficacy optimization model.
Because the anticoagulant drug effect optimization model consists of two parts, namely an initial drug effect optimization model and an adjusted drug effect optimization model, the anticoagulant drug effect optimization model can be respectively substituted into a specific model according to the test item data of the patient in different treatment stages to obtain the medication advice information of the patient in the treatment time period. Namely, if the patient is in the initial treatment stage, the test item data to be predicted is input into the initial pharmacodynamic optimization model, and an initial pharmacodynamic optimization scheme is obtained. And if the patient is in the treatment adjusting stage, inputting the data of the test item to be predicted into the adjustment efficacy optimization model to obtain an adjustment efficacy optimization scheme.
As shown in fig. 2, a schematic structural diagram of an apparatus for establishing an anticoagulant drug efficacy optimization model according to an embodiment of the present invention is provided, and the apparatus is used for implementing the method for establishing an anticoagulant drug efficacy optimization model described in each of the foregoing embodiments. Therefore, the description and definition of the method in the foregoing embodiments may be used for understanding the execution modules in the embodiments of the present invention.
The device includes: a sample acquisition module 201, a pre-processing module 202 and a model building module 203. Wherein,
the system comprises a sample acquisition module 201, a drug effect optimization sample database and a test item analysis module, wherein each sample in the drug effect optimization sample database comprises medication order information when the blood coagulation function is normal and test item data corresponding to the medication order information;
the preprocessing module 202 is configured to perform deficiency value processing and statistical inspection on the drug effect optimization sample database to obtain a preprocessed drug effect optimization sample database;
and the model establishing module 203 is used for performing supervised learning by using an XGboost algorithm according to the preprocessed drug effect optimization sample database to construct an anticoagulant drug effect optimization model.
The device for establishing the anticoagulant drug effect optimization model provided by the embodiment of the invention adopts the machine learning XGboost algorithm to establish the anticoagulant drug effect optimization model, and the inspection and inspection data of a patient are input into the anticoagulant drug effect optimization model, so that the anticoagulant drug effect optimization method with higher practicability can be quickly obtained, and the individualized drug effect optimization is realized.
Based on the content of the foregoing embodiment, the sample obtaining module 201 is specifically configured to:
acquiring clinical data of a patient using an anticoagulant and performing data cleaning;
sorting the clinical data of the patients after data cleaning according to time, extracting medication order information with an INR value of 1.5-2.5 after the patients use the anticoagulant drugs for the first time for three days and corresponding test item results from the clinical data of the patients to form an initial scheme database;
and extracting the medication order information when the INR value is not between 2 and 3 and the dosage is adjusted until the blood coagulation function of the patient is recovered to be normal after the first anticoagulant administration for three days and all the test item results after the dosage is adjusted to form an adjustment scheme database.
Based on the content of the foregoing embodiment, the preprocessing module 202 is specifically configured to:
deleting the inspection item with the data deletion rate larger than a preset threshold value in the drug effect optimization sample database;
and (3) performing data distribution inspection on the remaining inspection item data by adopting a statistical method, screening out significant variables influencing the drug effect of the anticoagulant drug, and obtaining a preprocessed drug effect optimization sample database.
Based on the content of the foregoing embodiment, the model establishing module 203 is specifically configured to:
initializing XGboost algorithm parameters, wherein the algorithm parameters comprise: maximum decision tree number, learning rate, maximum rule layer depth, minimum gain value required by decision tree growth and decision tree complexity measurement parameters;
based on the initial scheme database, screening out important variables influencing the initial drug effect of the anticoagulant drug by using an XGboost algorithm, and constructing an initial drug effect optimization model according to the important variables influencing the initial drug effect of the anticoagulant drug;
based on the adjustment scheme database, screening out important variables influencing the adjustment effect of the anticoagulant drug by using an XGboost algorithm, and constructing an adjustment effect optimization model according to the important variables influencing the adjustment effect of the anticoagulant drug;
and combining the initial efficacy optimization model and the adjusted efficacy optimization model to generate an anticoagulant efficacy optimization model.
As shown in fig. 3, a schematic structural diagram of an electronic device provided in an embodiment of the present invention includes: a processor (processor)301, a memory (memory)302, and a bus 303;
the processor 301 and the memory 302 respectively complete communication with each other through a bus 303; the processor 301 is configured to call the program instructions in the storage 302 to execute the method for establishing the anticoagulant drug efficacy optimization model provided in the above embodiments, for example, including: establishing a drug effect optimization sample database, wherein each sample in the drug effect optimization sample database comprises medication advice information when the blood coagulation function is normal and test item data corresponding to the medication advice information; carrying out deletion value processing and statistical inspection on the pesticide effect optimization sample database to obtain a pretreated pesticide effect optimization sample database; and carrying out supervised learning by using an XGboost algorithm according to the preprocessed drug effect optimization sample database to construct an anticoagulant drug effect optimization model.
An embodiment of the present invention further provides a non-transitory computer-readable storage medium, where the non-transitory computer-readable storage medium stores computer instructions, where the computer instructions cause a computer to execute the method for establishing an anticoagulant drug efficacy optimization model provided in the foregoing embodiment, for example, the method includes: establishing a drug effect optimization sample database, wherein each sample in the drug effect optimization sample database comprises medication advice information when the blood coagulation function is normal and test item data corresponding to the medication advice information; carrying out deletion value processing and statistical inspection on the pesticide effect optimization sample database to obtain a pretreated pesticide effect optimization sample database; and carrying out supervised learning by using an XGboost algorithm according to the preprocessed drug effect optimization sample database to construct an anticoagulant drug effect optimization model.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods of the various embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for establishing an anticoagulant drug effect optimization model is characterized by comprising the following steps:
establishing a drug effect optimization sample database, wherein each sample in the drug effect optimization sample database comprises medication advice information when the blood coagulation function is normal and test item data corresponding to the medication advice information;
carrying out deletion value processing and statistical inspection on the pesticide effect optimization sample database to obtain a pretreated pesticide effect optimization sample database;
and carrying out supervised learning by using an XGboost algorithm according to the preprocessed drug effect optimization sample database to construct an anticoagulant drug effect optimization model.
2. The method according to claim 1, wherein the step of establishing a database of pharmacodynamic optimization samples comprises:
acquiring clinical data of a patient using an anticoagulant and performing data cleaning;
sorting the clinical data of the patients after data cleaning according to time, extracting medication order information with an INR value of 1.5-2.5 after the patients use the anticoagulant drugs for the first time for three days and corresponding test item results from the clinical data of the patients to form an initial scheme database;
and extracting the medication order information when the INR value is not between 2 and 3 and the dosage is adjusted until the blood coagulation function of the patient is recovered to be normal after the first anticoagulant administration for three days and all the test item results after the dosage is adjusted to form an adjustment scheme database.
3. The method according to claim 1, wherein the step of performing deficiency value processing and statistical testing on the pharmacodynamic optimization sample database to obtain a preprocessed pharmacodynamic optimization sample database comprises:
deleting the inspection item with the data deletion rate larger than a preset threshold value in the drug effect optimization sample database;
and (3) performing data distribution inspection on the remaining inspection item data by adopting a statistical method, screening out significant variables influencing the drug effect of the anticoagulant drug, and obtaining a preprocessed drug effect optimization sample database.
4. The method according to claim 2, wherein the step of constructing an anticoagulant drug efficacy optimization model by performing supervised learning with an XGboost algorithm according to the preprocessed drug efficacy optimization sample database specifically comprises:
initializing XGboost algorithm parameters, wherein the algorithm parameters comprise: maximum decision tree number, learning rate, maximum rule layer depth, minimum gain value required by decision tree growth and decision tree complexity measurement parameters;
based on the initial scheme database, screening out important variables influencing the initial drug effect of the anticoagulant drug by using an XGboost algorithm, and constructing an initial drug effect optimization model according to the important variables influencing the initial drug effect of the anticoagulant drug;
based on the adjustment scheme database, screening out important variables influencing the adjustment effect of the anticoagulant drug by using an XGboost algorithm, and constructing an adjustment effect optimization model according to the important variables influencing the adjustment effect of the anticoagulant drug;
and combining the initial efficacy optimization model and the adjusted efficacy optimization model to generate an anticoagulant efficacy optimization model.
5. An establishing device of an anticoagulant drug effect optimization model is characterized by comprising:
the system comprises a sample acquisition module, a drug effect optimization sample database and a drug effect optimization analysis module, wherein each sample in the drug effect optimization sample database comprises medication advice information when the blood coagulation function is normal and test item data corresponding to the medication advice information;
the preprocessing module is used for carrying out deletion value processing and statistical inspection on the drug effect optimization sample database to obtain a preprocessed drug effect optimization sample database;
and the model establishing module is used for carrying out supervised learning by utilizing an XGboost algorithm according to the preprocessed drug effect optimization sample database to establish an anticoagulant drug effect optimization model.
6. The apparatus of claim 5, wherein the sample acquisition module is specifically configured to:
acquiring clinical data of a patient using an anticoagulant and performing data cleaning;
sorting the clinical data of the patients after data cleaning according to time, extracting medication order information with an INR value of 1.5-2.5 after the patients use the anticoagulant drugs for the first time for three days and corresponding test item results from the clinical data of the patients to form an initial scheme database;
and extracting the medication order information when the INR value is not between 2 and 3 and the dosage is adjusted until the blood coagulation function of the patient is recovered to be normal after the first anticoagulant administration for three days and all the test item results after the dosage is adjusted to form an adjustment scheme database.
7. The apparatus of claim 5, wherein the preprocessing module is specifically configured to:
deleting the inspection item with the data deletion rate larger than a preset threshold value in the drug effect optimization sample database;
and (3) performing data distribution inspection on the remaining inspection item data by adopting a statistical method, screening out significant variables influencing the drug effect of the anticoagulant drug, and obtaining a preprocessed drug effect optimization sample database.
8. The apparatus of claim 6, wherein the model building module is specifically configured to:
initializing XGboost algorithm parameters, wherein the algorithm parameters comprise: maximum decision tree number, learning rate, maximum rule layer depth, minimum gain value required by decision tree growth and decision tree complexity measurement parameters;
based on the initial scheme database, screening out important variables influencing the initial drug effect of the anticoagulant drug by using an XGboost algorithm, and constructing an initial drug effect optimization model according to the important variables influencing the initial drug effect of the anticoagulant drug;
based on the adjustment scheme database, screening out important variables influencing the adjustment effect of the anticoagulant drug by using an XGboost algorithm, and constructing an adjustment effect optimization model according to the important variables influencing the adjustment effect of the anticoagulant drug;
and combining the initial efficacy optimization model and the adjusted efficacy optimization model to generate an anticoagulant efficacy optimization model.
9. An electronic device, comprising:
at least one processor; and
at least one memory communicatively coupled to the processor, wherein:
the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the method of any of claims 1 to 4.
10. A non-transitory computer-readable storage medium storing computer instructions that cause a computer to perform the method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811045577.9A CN109243546B (en) | 2018-09-07 | 2018-09-07 | Method and device for establishing anticoagulant drug effect optimization model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811045577.9A CN109243546B (en) | 2018-09-07 | 2018-09-07 | Method and device for establishing anticoagulant drug effect optimization model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109243546A true CN109243546A (en) | 2019-01-18 |
| CN109243546B CN109243546B (en) | 2020-10-27 |
Family
ID=65067414
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811045577.9A Active CN109243546B (en) | 2018-09-07 | 2018-09-07 | Method and device for establishing anticoagulant drug effect optimization model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109243546B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110010252A (en) * | 2019-04-01 | 2019-07-12 | 上海交通大学医学院附属新华医院 | Warfarin dosage prediction technique and device |
| CN110021397A (en) * | 2019-02-01 | 2019-07-16 | 捷普科技（上海）有限公司 | Method and storage medium based on human body physiological parameter prediction dosage |
| CN111210890A (en) * | 2020-02-14 | 2020-05-29 | 成都木老仁康软件信息有限公司 | Anticoagulation pharmacy monitoring management method based on clinical data |
| CN111833985A (en) * | 2019-04-17 | 2020-10-27 | 复旦大学附属中山医院 | Insulin dosage form selection and dosage adjustment method and system |
| CN112466416A (en) * | 2020-11-03 | 2021-03-09 | 北京科技大学 | Material data cleaning method combined with prior knowledge of nickel-based alloy |
| CN116130117A (en) * | 2022-12-12 | 2023-05-16 | 海南省人民医院 | Access database-based method and device for realizing administration of anticoagulant drugs |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778042A (en) * | 2017-01-26 | 2017-05-31 | 中电科软件信息服务有限公司 | Cardio-cerebral vascular disease patient similarity analysis method and system |
| CN106919804A (en) * | 2017-03-22 | 2017-07-04 | 李学明 | Medicine based on clinical data recommends method, recommendation apparatus and server |
| CN108257675A (en) * | 2018-02-07 | 2018-07-06 | 平安科技（深圳）有限公司 | Chronic obstructive pulmonary disease onset risk Forecasting Methodology, server and computer readable storage medium |
| CN108367161A (en) * | 2017-06-05 | 2018-08-03 | 西安大医数码科技有限公司 | Radiotherapy system, data processing method and storage medium |
-
2018
- 2018-09-07 CN CN201811045577.9A patent/CN109243546B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778042A (en) * | 2017-01-26 | 2017-05-31 | 中电科软件信息服务有限公司 | Cardio-cerebral vascular disease patient similarity analysis method and system |
| CN106919804A (en) * | 2017-03-22 | 2017-07-04 | 李学明 | Medicine based on clinical data recommends method, recommendation apparatus and server |
| CN108367161A (en) * | 2017-06-05 | 2018-08-03 | 西安大医数码科技有限公司 | Radiotherapy system, data processing method and storage medium |
| CN108257675A (en) * | 2018-02-07 | 2018-07-06 | 平安科技（深圳）有限公司 | Chronic obstructive pulmonary disease onset risk Forecasting Methodology, server and computer readable storage medium |
Non-Patent Citations (4)
| Title |
|---|
| 张洪侠: "基于XGBoost算法的2型糖尿病精准预测模型研究", 《中国实验诊断学》 * |
| 张福康: "《医院常用药品处方集》", 31 December 2012, 东南大学出版社 * |
| 李晓琪: "适合中国人群抗凝治疗的华法林药物基因组学剂量预测模型建立和临床应用研究", 《中国优秀硕士学位论文全文数据库》 * |
| 贾文慧 等: "基于XGBoost模型的股骨颈骨折手术预后质量评分预测", 《太原理工大学学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110021397A (en) * | 2019-02-01 | 2019-07-16 | 捷普科技（上海）有限公司 | Method and storage medium based on human body physiological parameter prediction dosage |
| CN110010252A (en) * | 2019-04-01 | 2019-07-12 | 上海交通大学医学院附属新华医院 | Warfarin dosage prediction technique and device |
| CN111833985A (en) * | 2019-04-17 | 2020-10-27 | 复旦大学附属中山医院 | Insulin dosage form selection and dosage adjustment method and system |
| CN111210890A (en) * | 2020-02-14 | 2020-05-29 | 成都木老仁康软件信息有限公司 | Anticoagulation pharmacy monitoring management method based on clinical data |
| CN112466416A (en) * | 2020-11-03 | 2021-03-09 | 北京科技大学 | Material data cleaning method combined with prior knowledge of nickel-based alloy |
| CN112466416B (en) * | 2020-11-03 | 2024-04-12 | 北京科技大学 | Material data cleaning method combining nickel-based alloy priori knowledge |
| CN116130117A (en) * | 2022-12-12 | 2023-05-16 | 海南省人民医院 | Access database-based method and device for realizing administration of anticoagulant drugs |
| CN116130117B (en) * | 2022-12-12 | 2023-11-03 | 海南省人民医院 | Access database-based method and device for realizing administration of anticoagulant drugs |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109243546B (en) | 2020-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109243546B (en) | Method and device for establishing anticoagulant drug effect optimization model | |
| CN109243620B (en) | Drug effect optimization method and device based on blood concentration monitoring | |
| CN111613289B (en) | Individuation medicine dosage prediction method, device, electronic equipment and storage medium | |
| US20220254493A1 (en) | Chronic disease prediction system based on multi-task learning model | |
| CN113270203A (en) | Drug dose prediction method, device, electronic device and storage medium | |
| CN109741826B (en) | Anesthesia evaluation decision tree construction method and equipment | |
| WO2021151327A1 (en) | Triage data processing method and apparatus, and device and medium | |
| CN111243736A (en) | Survival risk assessment method and system | |
| CN109273097B (en) | Automatic generation method, device, equipment and storage medium for pharmaceutical indications | |
| CN109859815A (en) | A kind of adjuvant drug decision-making technique and intelligent adjuvant drug system | |
| CN109994179B (en) | Method and device for administering vancomycin | |
| CN113053535A (en) | Medical information prediction system and medical information prediction method | |
| CN111291131A (en) | Data processing method, data processing device, storage medium and electronic equipment | |
| CN112967803A (en) | Early mortality prediction method and system for emergency patients based on integrated model | |
| CN111785366A (en) | Method and device for determining patient treatment scheme and computer equipment | |
| CN113744832B (en) | Intelligent decision-making and quality control system for continuous kidney substitution therapy | |
| CN111062511A (en) | Aquaculture disease prediction method and system based on decision tree and neural network | |
| CN115966314B (en) | Data processing method and device, electronic equipment and storage medium | |
| CN109712715A (en) | A kind of scoring processing method and processing device of physiological data | |
| CN116313086A (en) | Sub-health prediction model construction method, device, equipment and storage medium | |
| CN111462914A (en) | Entity linking method and device | |
| CN116779156A (en) | Construction method of postoperative index anomaly prediction system and postoperative risk prediction equipment | |
| TWI792333B (en) | Prediction method and system of low blood pressure | |
| CN115295115A (en) | Sodium valproate blood concentration prediction method and device based on deep learning | |
| CN113057588A (en) | Disease early warning method, device, equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |